1. Field of the Invention
The present invention relates to an image searching device configured to search an image to be used from a database storing images of parts created on the basis of a three-dimensional model including multiple parts, an image searching method for searching the image, and a program used for searching the image.
2. Description of the Related Art
Recent improvements in computer performance and graphic processing technology have led to rapidly expanded use of three-dimensional models created by CAD (Computer-Aided Design) and CG (Computer Graphics). The three-dimensional models thus created are often used to create an image that can be viewed stereoscopically. In particular, the three-dimensional models are widely used in manufacturing industrial products. For example, the three-dimensional models are used to create an image to display a product model in a parts catalog, a service manual, and so forth. Further, contents including such an image are used as an electric medium.
Many industrial products, such as machines and electronic devices, include multiple parts. Therefore, a part included in the product is often required to be easily recognized in a full image of the entire product.
Conventionally, the part has been recognized by specifying a position or a region of the part in three dimensions on a screen displaying the three-dimensional model of the product, to extract the part on the basis of its geometric relations with other parts.
The above-described method for extracting the part is applied to a method disclosed, for example, in published unexamined Japanese Patent Application No. (hereinafter referred to as JP-A-) H09-190456. In a CAD system installing data on a three-dimensional product model, a closed space having a height, a width, and a depth is specified in three dimensions to extract a part belonging to the closed space thus specified. Further, JP-A-2006-039872 discloses a method in which a reference position, a reference direction, and a distance from the reference position in three dimensions are identified as geometric search criteria by a user using a stylus to extract a part that satisfies the geometric search criteria.
Thus, in both of the above-described methods, the three-dimensional space to be searched is specified by the user to extract the part. In other words, the user is required to identify the position of the part to be extracted in three dimensions.
Moreover, the user is required to perform a complicated operation to input data to specify the part to be extracted. Further, because geometric relations with other parts are calculated each time a part is extracted, a processing load increases as the number of parts to be extracted increases. Consequently, a lower performance processor cannot perform the required larger amount of calculation due to lack of processing speed.
To solve such problems, JP-A-2007-042077 discloses a method in which a two-dimensional closed region is specified by the user on a screen displaying a two-dimensional image of an entire model created on the basis of a three-dimensional model to extract a part belonging to the two-dimensional closed region thus specified. In the above-described method, multiple parts belonging to the two-dimensional closed region are displayed on the screen such that the user can select a target part from the multiple parts. Further, the target part selected by the user is highlighted in the two-dimensional image of the entire model.
In the method disclosed in JP-A-2007-042077, when the target part is searched, the closed region is specified in the two-dimensional image of the entire model displayed on the screen in order to limit the number of parts to be searched to the parts belonging to the closed region thus specified. In addition, a list of single images of the parts belonging to the closed region is displayed on the screen. When a single image of a part in the list is temporarily selected by the user, such a part is highlighted in the two-dimensional image of the entire model on the screen so that the user can visually confirm whether or not the part temporarily selected is a target part.
However, in a case in which the part temporarily selected is too small relative to the two-dimensional image of the entire model, it is difficult for the user to visually confirm the part even when the part is highlighted, resulting in lower work efficiency of the user.